UM 3DT Review and Measurements

A few months back, I posted some thoughts on why I felt balanced armature IEMs suck: https://www.superbestaudiofriends.org/index.php?threads/why-balanced-armature-drivers-suck.11485/

I think or at least I would hope that readers understand where I was coming from; because after all, the CFA Ara and OG Solaris remain my go to IEMs - and these two IEMs use balanced armatures at least in some spots (I'm bummed because I lost my Aras, or I think they may have been stolen by the guy who installed our blinds in the house). Here at SBAF, discussions must all be taken into context. We're not afraid to rip apart the stuff we like because we can see something potentially better. Our ego's aren't wrapped into our gear.

In that thread, it was suggested that instead of the DD/BA hybrid approaches which are always problematic because of discontinuities between the driver types, a DD/DD approach akin to mutli-driver speakers could be the ticket. In theory, perhaps a multi-driver DD would fix the downsides that seem to come with every DD IEM: peaky highs.

UM has taken this approach with the 3DT. The 3DY uses x2 DD bass drivers and another x1 DD driver for the mids and highs. Does the 3DT bring us the best of DD tech without the downsides of DD?

Unfortunately, the answer is no.

UM 3DT IEM
Frequency Response

 

I couldn't find any 3M micropore tape at the drugstore yesterday so I will have to wait under the Amazon order comes. In the meantime, I will try some mask material and report back.

No doubt the UM 3DT is going to be compared to the FDX1 because both are DD and both are in the $300 (USD) range, so let's start there.

I preferred the FDX1 with the blue filter which was the most aggressive in pushing that upper midrange bump down. I know it's a matter of semantics maybe, but I wouldn't have categorized FDX1's upper mids as peaky, but rather as elevated or bumped. The 3DT sounds not just elevated, but also peaky, implying several high Q peaks or unevenness. The FDX1 has a smoother response. Crucially, the FDX1 not as hot on the upper fridges of the upper mids and lower mids.

UM 3DT vs Drop FDX1 (both stock, or non-colored nozzle for FDX1)
Frequency Response


Where the 3DT subjectively beats the FDX1 in the lows. The 3DT seems to be a little less physicality compared to the FDX1, but the FDX1 can lay the bass down a bit thick at times. The 3DT bass also sounds quite nimble or "fast" for a DD. With respect to transients, the UM 3DT reminds me of the DD drivers used in the CFA hybrids (their timbre and texture is different).

Finally last few points: The 3DT still doesn't quite sound as cohesive throughout the audio band compared to the FDX1. The FDX1 sounds more resolving, higher-fidelity, than the 3DT. There's kind of a "weaksauce" (relative) sound to the 3DT which is hard for me to explain - other than it sounds like tube families that I do not like, e.g. 12AX7. I'll take distortion measurements and report back comparing to FDX1. I'm betting there is something there.

 

Ignore the 200Hz 3rd harmonic spikes (green line) in the measurements below.

UM 3DT
Distortion (absolute) at ~104db


Drop FDX1
Distortion (absolute) at ~104db


Overall H2 is similar around -70db in the lower mids. FDX1 has a little higher H2 starting below 80Hz. FDX1 also has slightly higher H2 in the upper mids.

However, I think the main difference is with the higher orders. H4 on the FD1X stays low throughout, a slight rise below 80Hz. Similar for H3, gradual rise below 100Hz - but with a narrow bump at 3.4kHz. But otherwise from 100 to 2.6kHz, the higher orders remain non-existent on the FD1X.

The same cannot be said for the UM 3DT. There's a nice low spot between 300 and 1.6kHz. From up 1.6kHz, there's a steady rise that keeps going up. Taking into account the H3 200Hz artifact, H3 on the 3DT is higher too below 300Hz. There's also a H2 and H3 distortion peak just past 5kHz.

 

To be clear, more resolving of microdetail in the amplitude domain in the highs. With transients, time domain, the FDX1 is sluggish. The bass is smeared. The 3DT has better definition, probably because of a mid-treble peak (not totally sure it's there per the FR), but the plankton seems fake to me, that is the Photoshop sharpening effect.

I'm also wondering if there is an ultrasonic peak somewhere. The timbre of the highs on the 3DT is a bit unnatural to me. I wonder if it's related to that H2 spike just past 10kHz, and also that H2/H3 spike just past 5kHz.

Now I have yet to mod the 3DT.

 

The answer is simple: it's because people keep sidegrading.

https://www.superbestaudiofriends.o...d-iem-hate-confusion.11787/page-2#post-369349

85% of the effort for product development goes toward getting it right, final tweaks, just before production. Most people don't realize how many prototypes from Eddie Current, Schiit, Drop, etc. get left in the dustbin. For every one product that makes it to production there are at least three protos that don't, and five ideas, proof-of-concepts, that got killed early on.

True stories:

"Hey Marv, check this out. I want to hear your take on this."
"It sounds like ummm, how do I put this..."
"Like shit?"
"Yeah"

"Hey Marv, check these IEMs out"
"Why do the left and right channels sound different, which is it supposed to be?"
"We'll get you another good sample"
"Why does this good sample sound yet different from the original one?"
"Well get you another three samples"
"WTF, why does everything sound different?"
"Oh shit"
"Just price them at $39"
"Good idea"

I'd be willing to bet that most of these IEMs manufacturers just kind of say to themselves "f**k it, good enough, write dumb marketing document highlighting our goofy tech (meta materials, meta drivers, meta housings, number of drivers, etc.) send a few free ones to shills on HF, people will buy it anyway".

I mean, really, I have to find micropore tape to fix a fucked sounding IEM? I'm tired of having to do stuff that the manufacturer should have done themselves in a more elegant manner in the first place (Helmholtz resonators, special chambers and sound guides, Holtzman effect, Axlotil tanks, etc.)

 

I don't mind minor tweaks, e.g. tape mod on Grado bowl pads. However, finding the right tape, cutting it, and dicking with tiny IEM shit, that takes time and is annoying. At least Grado listened and provided F-pads for the SR325X and SR225X. And then there's the GAudio Nair, where I did absolutely nothing, nada, zero effort, and they sounded wonderful and right.

 

Yeah, that's why everyone on Changstar/SBAF shit on or ignored them for past 10 years and preferred the Joe Grado HP1000 from the early 80s.

---

UM 3DT
Frequency Response
RED = Stock
CYN = x2 layers outer mask material
ORA = x1 layer out and x1 layer middle mask material


UM 3DT
x1 layer outer and x1 layer middle mask material
Distortion at ~104db


Much better now. The 5kHz peak looks to have been eliminated. Not only that, the 5kHz distortion is gone. The distortion characteristics actually look very nice with H2 being dominate, and H3 and H4 being well below H2. There is a slightly rise in H3 in the upper mids, but nothing's perfect.

 

Now we are making some progress. I took two layers of the three layer mask material sandwidth. I'll try the micropore when it comes in since this x2 layer sandwich could be a bit on the aggressive side with making the tonal response more amenable. Then again, I don't think the micropore will be enough for me - it doesn't kill that 5kHz peak - so I could be stuck in a bad place.

The main downside with shoving stuff in front of the nozzle is attenuation of the last octave and some killing of liveliness (perhaps we can call this the Aeon effect). This is any different here, but the effect isn't severe. It's like most DD IEMs don't have plenty of air anyway. Even without mods, the lack of air and shimmer with metal percussion was notable, especially compared to a proper headphone HD6**. The sound is a bit stuffier though compared to IEMs that just sound good without having to do tape or material mods in such a heavy handed way.

With respect to subjective tonal response, the HD600 and CFA OG Solaris, even with their imperfections, sound so much smoother and grown up. The 3DT sounds less refined and rather sizzly. The timbre is much better with the modded 3DT, however, there's still that spotlit peak that won't go away. Compared to the OG Solaris, it's like pick your poison: some BA timbre vs. spot-lit sizzle. I'd call it even. The DD goodness still comes with its drawbacks: peakage. Rawer BA timbre still comes with its advantages: clarity and transients. Although it's probably not fair to compare with the CFAs since their BA implementations are among the better ones (and also in another price class).

The 3DT does have some "plankton" that only DDs seem to be able to pick up. It's not on IE800 or Dunu Zen (I have not heard Pro yet) level. It's most certainly not on HD6** level. But it pips any BA, including any high-end ones. I'd say about on par with Grado SR225's ability to resolve low-level information which isn't bad - quite good really (This is really for you @crenca, IEMs are not ready for you).

Component chain: Yggdrasil LIM -> Magni 3+
Music used: Daft Punk, Alice in Chains Unplugged, lots of old big band stuff from Eartha Kitt and Bing Crosby.

 

In a nutshell:

• The price is right, especially with the recent price reduction.
• Conditionally highly recommended, provided you are willing to put in some elbow grease. Difficult highs. timbre, and distortion are largely addressed with tape mod.
• Doubtful everything can be fixed. Mid-treble peakage cannot be totally fixed. There are downsides with taping stuff over the nozzle, stuffier sound, and attenuation of last octave. Some experimentation for ideal material to suit personal tastes and find the right trade-offs may be necessary.
• Nice DD timbre, but DD downsides (treble peakage) not fully addressed
• Good resolution considering price.

 

The FDX1 seems to prefer desktop amps that can cut through, otherwise the lows are a mushy mess. Also, desktop gear is more resolving, which the FDX1 can take advantage.

x2 7mm DD for the lows.
x1 10mm carbon nanotube DD for the highs.

The Carbon Nanotube stuff is the latest crazy, now that folks realized EST sounded like shit.

UM 3DT
Impedance and Electrical Phase

 

Yeah, synergy for sure. I like the FDX from more neutral amps Magni 3+ to organic sounding portable players such as the Sony or M6.

--

Anyway, I got the 3M micropore tape. These are the results with FULL coverage of the nozzle.

UM 3DT
Frequency Response
GRY= Stock
YEL = 3M micropore tape


This sort of gives it an Audeze type frequency response.. TBH, I still prefer the FDX1. The FDX1 isn't perfect, thick murky lows that obscure timbre, emphasized upper mids that extends up a bit past pinna gain; but the showstopper for me with respect to the 3DT is the grating splashy treble timbre, this hard to describe weaksauce sound, and lower resolution mids and highs.